Supplementary webappendixThis webappendix formed part of the original submission and has been peer reviewed. We post it as supplied by the authors.

Supplement to: Vos SJB, Xiong C, Visser PJ, et al. Preclinical Alzheimer’s disease and its outcome: a longitudinal cohort study. Lancet Neurol 2013; published online Sept 4. http://dx.doi.org/10.1016/S1474-4422(13)70194-7.

  1

Preclinical Alzheimer’s disease and its outcome: a longitudinal cohort study

Stephanie J.B. Vos MSc, Chengjie Xiong PhD, Pieter Jelle Visser MD, Mateusz S. Jasielec MS, Jason

Hassenstab PhD, Elizabeth A. Grant PhD, Nigel J. Cairns PhD, John C. Morris MD, David M. Holtzman

MD, Anne M. Fagan PhD

Supplemental Table 1. Baseline demographics of the independent CDR 0·5 symptomatic AD sample Supplemental Table 2. Overview of concepts for mild cognitive impairments Supplemental Table 3. Exact p-values for the pairwise comparisons of Table 1 Supplemental Table 4. Distribution of preclinical AD stages Supplemental Table 5. Preclinical AD and its outcome according to age and APOE genotype Supplemental Figure 1. Annual rate of change in CDR-SB and MMSE by preclinical AD stage Supplemental Figure 2. Individual cognitive trajectories for CDR-SB and MMSE by preclinical AD stage Supplemental Text 1. Cohort informati0n Supplemental Table 6. Preclinical AD and its outcome using different classification approaches Supplemental Table 7. Prediction of the unclassified group for CDR0·5 symptomatic AD, mortality, and annual cognitive decline

  2

Supplemental Table 1. Baseline demographics of the independent CDR 0·5 symptomatic AD sample (N=74)

Results are mean (SD) or number (%). Episodic memory is a composite score of the Associate Learning Test, Logical Memory Test, and Selective Reminding Test. AD=Alzheimer’s disease, APOE=Apolipoprotein E, MMSE=Mini-Mental State Examination (range 0-30, with 30 as the best score), CDR-SB=Clinical Dementia Rating scale Sum of Boxes (range 0-16, with 0 as the best score), Aβ=beta amyloid, p-tau=phosphorylated tau, t-tau=total tau.

Age 74·5 (5·4)

Female, n 36 (49%)

Education, y 14·5 (3·1)

Ethnic origin, n

White 71 (96%)

African-American 2 (3%)

Native Hawaiian and Pacific Islander 1 (1%)

APOE-4+, n 51 (69%)

MMSE 25·9 (3·0)

CDR-SB 2·5 (1·0)

Episodic memory, z-score -1·6 (1·0)

Aβ1-42, pg/mL 424 (219)

T-tau, pg/mL 607 (293)

P-tau181, pg/mL 97 (48)

  3

Supplementary Table 2. Overview of concepts for mild cognitive impairments Construct Definition Psychometric test

impairment Functional impairment Clinical dementia

diagnosis according to DSM IV or ICD-10

Biomarker status

CDR 0·5 symptomatic AD1

Defined cut-offs not utilized

Very mild to mild change in daily functioning in memory and at least 1 non-memory domain

Some Not needed

Amnestic MCI2

Yes, in memory domain Subjective report of cognitive decline

No Not needed

MCI due to AD3

Yes, in any cognitive domain

Cognitive concern reflecting a change in cognition

No Abnormal

Prodromal AD4

Yes, in memory domain Complaints of memory decline No Abnormal

References 1. Morris JC. The Clinical Dementia Rating (CDR): current version and scoring rules.

Neurology 1993; 43: 2412–14. 2. Petersen RC. Mild cognitive impairment as a diagnostic entity. Journal of Internal

Medicine 2004; 256: 183–94. 3. Albert MS, DeKosky ST, Dickson D, Dubois B, Feldman HH, Fox NC, et al. The

diagnosis of mild cognitive impairment due to Alzheimer’s disease: Recommendations from the National Institute on Aging and Alzheimer’s Association workgroup. Alzheimers Dement 2011; 7: 270–9.

4. Dubois B, Feldman HH, Jacova C, Dekosky ST, Barberger-Gateau P, Cummings J, et al. Research criteria for the diagnosis of Alzheimer's disease: revising the NINCDS-ADRDA criteria. Lancet Neurol 2007; 6: 734–46.

  4

Supplemental Table 3. Exact p-values for the pairwise comparisons of Table 1.

Results are p-values of pairwise comparisons of Table 1. Aβ= amyloid-beta, APOE=Apolipoprotein E, CDR-SB=Clinical Dementia Rating scale Sum of Boxes, MMSE=Mini-Mental State Examination, p-tau=phosphorylated tau, SNAP=Suspected Non-Alzheimer Pathophysiology, t-tau= total tau.

Normal vs. stage 1

Normal vs. stage 2

Normal vs. stage 3

Normal vs. SNAP

Normal vs. unclassified

Stage 1 vs. stage 2

Stage 1 vs. stage 3

Stage 1 vs. SNAP

Stage 1 vs. unclassified

Stage 2 vs. stage 3

Stage 2 vs. SNAP

Stage 2 vs. unclassified

Stage 3 vs. SNAP

Stage 3 vs. unclassified

SNAP vs. unclassified

Age 0·0007 0·0003 <0·0001 0·0003 <0·0001 0·6531 0·0146 0·8184 0·0084 0·0393 0·4849 0·0254 0·0074 0·9222 0·0039 Female 0·3678 0·4333 0·0859 0·6665 0·1451 0·1763 0·2501 0·2482 0·3864 0·0465 0·6756 0·0778 0·0627 0·7855 0·1060 Education 0·0004 0·7729 0·2942 0·7115 0·3502 0·0127 0·0037 0·0033 0·0044 0·2665 1·000 0·3138 0·2329 0·9123 0·2775

APOE-4 0·0059 0·0045 0·0025 0·3787 0·7804 0·7731 0·1601 0·0744 0·1008 0·2380 0·0507 0·0758 0·0125 0·0173 0·4943

MMSE 0·0101 0·2195 <0·0001 0·5203 0·0101 0·3397 0·0024 0·0646 0·3532 0·0003 0·5019 0·1159 <0·0001 0·0791 0·0308 CDR-SB 0·3347 0·8380 0·1167 0·8884 0·7059 0·5697 0·3502 0·3239 0·8482 0·1961 0·7719 0·8300 0·1134 0·3621 0·6642 Episodic memory

0·8437 0·2302 <0·0001 0·6128 <0·0001 0·2409 <0·0001 0·5646 <0·0001 <0·0001 0·4567 <0·0001 <0·0001 0·0671 <0·0001

Aβ1-42 <0·0001 <0·0001 <0·0001 0·0396 0·1847 0·8954 0·5672 <0·0001 <0·0001 0·6425 <0·0001 <0·0001 <0·0001 <0·0001 0·0209 T-tau 0·5470 <0·0001 <0·0001 <0·0001 0·5634 <0·0001 <0·0001 <0·0001 0·8438 0·3126 0·0004 <0·0001 0·0005 <0·0001 <0·0001 P-tau181 0·9111 <0·0001 <0·0001 <0·0001 0·1929 <0·0001 <0·0001 <0·0001 0·2537 0·7014 0·0638 <0·0001 0·0955 <0·0001 <0·0001 Follow-up 0·6396 0·7969 0·1251 0·9818 0·8928 0·5625 0·0934 0·6575 0·8906 0·2185 0·8249 0·7838 0·1415 0·2084 0·8876 Progression to CDR0·5

0·0077 0·0002 <0·0001 0·1338 0·0005 0·1575 0·0033 0·1770 0·1716 0·0641 0·0069 0·7962 0·0001 0·1875 0·0145

Mortality 0·0180 0·0604 0·0003 0·0673 0·2077 0·7250 0·0850 0·4808 0·7018 0·0623 0·7953 0·8893 0·0187 0·1436 0·9787

  5

Supplemental Table 4. Distribution of preclinical AD stages

Results are number (%) of participants identified in the different stages based on CSF markers with optimal Youden cut-offs: Abnormal CSF Aβ1-42 <459 pg/mL, t-tau >339 pg/mL, p-tau181 >67 pg/mL. Cognition was an episodic memory composite score of the Associate Learning Test, Logical Memory Test, and Selective Reminding Test, with a cut-off at the lowest 10th percentile: -1·25 SD. Aβ=beta amyloid, AD=Alzheimer’s disease, p-tau=phosphorylated tau, SNAP=Suspected Non-Alzheimer Pathophysiology, t-tau=total tau.

Stage CSF Aβ1-42 <459 pg/mL

CSF t-tau >339 pg/mL

CSF p-tau181 >67 pg/mL

Cognition <-1·25 SD

N (%) Overall N (%)

Normal group - - - - 129 (41·5) 129 (41·5)

Stage 1 + - - - 47 (15) 47 (15)

Stage 2 + + - - 8 (3) 36 (12)

+ - + - 5 (2)

+ + + - 23 (7)

Stage 3 + + - + 1 (0·5) 13 (4)

+ + + + 12 (3·5)

SNAP group - + - - 14 (4·5) 72 (23)

- + - + 2 (0·5)

- - + - 9 (3)

- + + - 45 (14·5)

- + + + 2 (0·5)

Unclassified - - - + 11 (3·5) 14 (4·5)

+ - - + 3 (1)

  6

Supplemental Table 5. Preclinical AD and its outcome according to age and APOE genotype

Data are baseline number (%) of participants with and without preclinical AD, and number (%) of these participants that progressed to CDR0·5 symptomatic AD by dichotomous age (A) and APOE genotype (B) grouping. AD=Alzheimer’s disease, APOE=Apolipoprotein E, CDR=Clinical Dementia Rating scale, SNAP=Suspected Non-Alzheimer Pathophysiology. *P<0·05 compared to age ≤72, ***p<0·001 compared to APOE-4-, based on Chi-squared tests for 2 by 2 tables.

A Baseline prevalence Progression to CDR≥0·5 symptomatic AD

Age ≤72 (n=169)

Age >72 (n=142)

Age ≤72 (n=6)

Age >72 (n=25)

No preclinical AD 125 (74%) 90 (64·5%)* 1 (1%) 9 (10%)

Normal group 90 (53·5%) 39 (27·5%) 0 (0%) 2 (5%)

SNAP group 31 (18·5%) 41 (29%) 1 (3%) 3 (7%)

Unclassified 4 (2%) 10 (7%) 0 (0%) 4 (40%)

Preclinical AD 44 (26%) 52 (36·5%) 5 (11%) 17 (33%)

Stage 1 25 (15%) 22 (15·5%) 1 (4%) 5 (23%)

Stage 2 17 (10%) 19 (13%) 2 (12%) 7 (37%)

Stage 3 2 (1%) 11 (8%) 2 (100%) 5 (45·5%)

B Baseline prevalence Progression to CDR≥0·5 symptomatic AD

APOE-4-

(n=205)

APOE-4+

(n=106)

APOE-4-

(n=18)

APOE-4+

(n=13)

No preclinical AD 158 (77%) 57 (53%)***

7 (4%) 3 (5%)

Normal group 97 (47%) 32 (30%) 2 (2%) 0 (0%)

SNAP group 50 (24·5%) 22 (21%) 2 (4%) 2 (9%)

Unclassified 11 (5·5%) 3 (2%) 3 (27%) 1 (33%)

Preclinical AD 47 (23%) 49 (47%) 11 (23%) 11 (22%)

Stage 1 25 (12%) 22 (21%) 2 (8%) 4 (18%)

Stage 2 18 (9%) 18 (17%) 6 (33%) 3 (17%)

Stage 3 4 (2%) 9 (9%) 3 (75%) 4 (44%)

  7

Supplemental Figure 1. Annual rate of change in CDR-SB and MMSE by preclinical AD stage Figure legend: Graphs show the estimated annual rate of change in CDR-SB (A) and MMSE (B), based on slopes according to each preclinical AD stage, corrected for age, gender, education, and APOE genotype. The black line represents participants in the normal group; light blue, stage 1; dark blue, stage 2; red, stage 3; and grey, SNAP. CDR-SB=Clinical Dementia Rating scale Sum of Boxes (range 0-18, with 0 as the best score), MMSE=Mini-Mental State Examination (range 0-30, with 30 as the best score).

  8

Supplemental Figure 2. Individual cognitive trajectories for CDR-SB and MMSE by preclinical AD stage Figure legend: Graphs show individual cognitive trajectories (blue) and the overall estimated annual rate of change (red) for CDR-SB (A) and MMSE (B) by preclinical AD stage. CDR-SB=Clinical Dementia Rating scale Sum of Boxes (range 0-18, with 0 as the best score), MMSE=Mini-Mental State Examination (range 0-30, with 30 as the best score), SNAP= Suspected Non-Alzheimer Pathophysiology.

  9

  10

Supplemental Text 1. Cohort information The overall rate of progression from CDR 0 to CDR 0.5 or greater in our sample was 10% (32/311) over an average of 4 years of follow-up. This percent is comparable to the Mayo Clinic Study of Aging (MCSA), a population based longitudinal program in which many participants have multiple imaging procedures and some have lumbar puncture. The overall rate of progression from cognitive normality to MCI in the MCSA sample was 20% (296/1450) with a median follow-up of 3·4 years, but the MCI construct in the MCSA program was unstable as 34% of MCI individuals reverted to cognitive normality at subsequent follow-up.1

Removing the MCSA individuals who later reverted to normal leaves 196 of the 1450 (13·5%) cognitively normal persons who developed MCI/symptomatic AD, very similar to our 10%.

References

1. Roberts RO, Geda YE, Knopman DS, et al. The incidence of MCI differs by subtype and is higher in men: the Mayo Clinic Study of Aging. Neurology 2012; 78: 342–51.

  11

Supplemental Table 6. Preclinical AD and its outcome using different classification approaches

A) Use of different CSF cutoffs; Aβ1-42 <500, t-tau <440, p-tau181 <78 pg/ml

B) Stage 3 defined as a score in the lowest 10th percentile of any cognitive domain

C) Participants with a baseline CDR-SB of 0·5 excluded (n=293)

 Results are number (%) for (A) previously applied CSF cut-offs of our center, (B) a different definition of stage 3 based on the episodic memory, semantic memory, working memory, and visuospatial composite score (described in Johnson et al 2009), and (C) only participants with CDR-SB=0 at baseline. AD=Alzheimer’s disease, CDR-SB=Clinical Dementia Rating scale Sum of Boxes (range 0-18, with 0 as the best score), CSF=cerebrospinal fluid, SNAP= Suspected Non-Alzheimer Pathophysiology.

Normal group Stage 1 Stage 2 Stage 3 SNAP group Unclassified group Proportion stages, n 131 (42%) 85 (27%) 28 (9%) 13 (4%) 37 (12%) 17 (6%) Progression to CDR≥0·5 symptomatic AD, n

2 (1.5%) 9 (11%) 7 (25%) 7 (54%) 2 (5%) 5 (29%)

Mortality, n 1 (1%) 7 (8%) 3 (11%) 4 (31%) 4 (11%) 1 (6%)

Normal group Stage 1 Stage 2 Stage 3 SNAP group Unclassified group Proportion stages, n 104 (33%) 37 (12%) 28 (9%) 21 (7%) 71 (23%) 50 (16%) Progression to CDR≥0·5 symptomatic AD, n

2 (2%) 6 (16%) 8 (29%) 8 (38%) 4 (6%) 4 (8%)

Mortality, n 0 (0%) 5 (14%) 1 (4%) 6 (29%) 4 (6%) 4 (8%)

Normal group Stage 1 Stage 2 Stage 3 SNAP group Unclassified group Proportion stages, n 123 (42%) 43 (15%) 34 (12%) 11 (4%) 69 (24%) 13 (4%) Progression to CDR≥0·5 symptomatic AD, n

2 (2%) 6 (14%) 9 (27%) 6 (55%) 4 (6%) 3 (23%)

Mortality, n 2 (2%) 5 (12%) 2 (6%) 4 (36%) 4 (6%) 0 (0%)

  12

Supplemental Table 7. Prediction of the unclassified group for CDR0·5 symptomatic AD, mortality, and annual cognitive decline  A) Prediction of the unclassified group for CDR0·5 symptomatic AD and mortality

Progression to CDR0·5 symptomatic AD Mortality risk

5-year progression rate

Uncorrected SHR (95% CI)

Difference compared to normal group

Corrected SHR (95% CI)

Difference compared to normal group

Uncorrected HR (95% CI)

Difference compared to normal group

Corrected HR (95% CI)

Difference compared to normal group

Normal group 2% Reference Reference Reference Reference

Unclassified group 34% 24·2 (4·7-125·5)

P=0·0002 16·4 (2·8-95·8)

P=0·0019 5·3 (0·5-59·9)

P=0·1749 2·8 (0·2-36·9)

P=0·4305

 Results are 5-year progression rate (cumulative incidence rate) to CDR0·5 symptomatic AD and associated subhazard ratio (SHR, 95% CI) for progression to CDR0·5 symptomatic AD calculated using a Fine & Gray subdistribution hazards model, and hazard ratio (HR, 95% CI) for mortality calculated using Cox regression analyses. Analyses are shown as both uncorrected and corrected for baseline age, gender, education, and APOE genotype. AD=Alzheimer’s disease, CDR=Clinical Dementia Rating scale, HR=Hazard Ratio, SHR=Subhazard Ratio, SNAP=Suspected Non-Alzheimer Pathophysiology.    B) Annual rate of change in CDR-SB and MMSE for the unclassified group  Stages Slope CDR-SB P-value slope Difference compared

to normal group Slope MMSE P-value slope Difference compared

to normal group

Normal group 0·03 (0·03) P=0·2661 Reference -0·01 (0·03) P=0·6673 Reference

Unclassified group 0·33 (0·08) P=0·0001 P=0·0007 -0·24 (0·10) P=0·0127 P=0·0258

 Data are slopes (SE) corrected for age, gender, education, and APOE genotype and comparison to other groups. AD=Alzheimer’s disease, CDR-SB=Clinical Dementia Rating scale Sum of Boxes (range 0-18, with 0 as the best score), MMSE=Mini-Mental State Examination (range 0-30, with 30 as the best score), SNAP=Suspected Non-Alzheimer Pathophysiology.